Charge pump regulation circuit to increase program and erase efficiency in nonvolatile memory

A charge pump circuit generates a charge pump output signal at a first node and is enabled by a control signal. A diode has an anode coupled to the first node and a cathode coupled to a second node. A current mirror arrangement sources a first current to the second node and sinks a second current from a third node. A comparator causes the control signal to direct the charge pump circuit to generate the charge pump output signal as having a voltage that ramps upwardly in magnitude (but negative in sign) if the voltage at the second node is greater than the voltage at the third node, and causes the control signal to direct the charge pump circuit to cease the ramping of the voltage of the charge pump output signal if the voltage at the second node is at least equal to the voltage at the third node.

Skip to: Description  ·  Claims  ·  References Cited  · Patent History  ·  Patent History
Description
RELATED APPLICATION

This application is a continuation-in-part of U.S. application patent Ser. No. 16/715,209, filed on Dec. 16, 2019, which claims priority to U.S. Provisional Application for Patent No. 62/795,853, filed Jan. 23, 2019, the contents of which are incorporated by reference in their entirety.

TECHNICAL FIELD

This disclosure is related to a charge pump and associated regulation circuitry for generating program and erase voltages for use in a non-volatile memory.

BACKGROUND

Non-volatile memory devices are used in applications where it is desirable to store information that is to be retained even when the device into which the non-volatile memory is incorporated is not powered. Generally, each non-volatile memory device includes a matrix of memory cells utilizing floating gate MOS storage transistors; each such floating gate MOS storage transistor has a threshold voltage that may be set (according to an electric charge in its floating gate) to different levels to represent different logic values.

For example, in Electrically Erasable and Programmable Read-Only Memories (EEPROMs) each floating gate MOS storage transistor may be both programmed (by injecting electric charge into its floating gate) or erased (by removing electric charge from its floating gate) individually thanks to the use of a set of MOS selection transistors that selectively apply the required voltages to the corresponding storage floating gate MOS transistor (with a complex structure that limits the capacity of the EEPROMs). On the other hand, flash memories have a simple structure that allows obtainment of very high capacities of hundreds of Gigabytes or even higher thanks to the grouping of the floating gate MOS transistors in sectors, each integrated in a common well of semiconductor material without a selection transistor, but with a structure that enables the erasing at a sector level instead of at an individual level.

In both cases, as well as in the case of any non-volatile memory, program and erase efficiency is a function of the electric field applied across the floating gate MOS storage transistors. To generate these electric fields, a high voltage is generated by a charge pump and applied to a word line (for program operations) or applied to a substrate line (for erase operations). Floating gate MOS storage transistors have parasitic diodes formed by the PN junction between their source and drain. During both the program and erase operations, these parasitic diodes are reverse biased.

Some non-volatile memory cells may be made using a triple well technology where an isolated Pwell structure is used to make an NMOS device. There are two extra parasitic diodes present in triple well technology: one PN junction (and thus parasitic diode) is between the isolated Pwell and the Nwell and another PN junction (and thus parasitic diode) is between the Nwell and the substrate. In any design, these junctions should be reverse biased so that proper functionality can be maintained during at least program and erase operations.

Due to process and temperature variations, some parasitic diodes of a non-volatile memory cell can be being electrically stressed, which could lead to breakdown, and improper function (or even damage) of the non-volatile memory cell.

In order to avoid such condition, it is possible to define rail voltages for these parasitic PN junctions in triple well technology. Normally the PN junction between an isolated Pwell and an Nwell is named as “Intrail” and the PN junction between the Nwell and the substrate is named as “Extrail”. Rail voltage limits can be defined for these PN junctions, which serve to change the design rule for the PN junctions and eventually changes the break-down voltage as well. For example, if a designer knows that maximum voltage across the PN junctions is going to be 15V, then a rail voltage of more than 15V will be defined to avoid break-down of these PN junctions. However, this leads to an increased area cost and potentially an increased production cost.

Therefore, to avoid the defining of over-estimated rail voltages and the associated area and production cost, a new regulation scheme for the charge pump circuits has been developed. An example charge pump and regulation circuit 50 is shown in FIG. 1. The charge pump and regulation circuit 50 includes a charge pump 52 that generates a charge pump output voltage OUT under control of the clock signals CLK and CLKN, which are logical inverts of each other. A resistive divider formed from resistors R1 and R2 is coupled in series between the output of the charge pump 52 and ground. Node N denotes the partitioning of the resistive divider. A comparator 54 has its non-inverting input coupled to receive a bandgap voltage VBG (or any reference voltage which is stable across different processes, temperature, and supply conditions), its inverting input coupled to the partitioning of the resistive divider at node N, and generates a control signal OSTOP at its output, and OSTOP can be either an oscillator or a clock control signal. The oscillator control signal OSTOP is an on/off enable signal and controls an oscillator 60 that generates the clock signals CLK and CLKN that in turn control the charge pump 52. When OSTOP is at a logic high, the oscillator 60 is enabled, and when OSTOP is at a logic low, the oscillator 60 is disabled.

In operation, the voltage divider serves to detect the output voltage OUT of the charge pump 52 at node N. The voltage at node N is then compared to the bandgap voltage VBG. If the voltage at node N is below the bandgap voltage VBG, the comparator 54 generates the oscillator control signal OSTOP at a logic high. When the oscillator control signal OSTOP is at a logic high, the oscillator 60 generates or outputs the clock signals CLK and CLKN to cause the charge pump 52 to ramp up the voltage of the charge pump output signal OUT. The voltage of the charge pump output signal OUT therefore continues to ramp up until the voltage at node N is equal to the bandgap voltage VBG, at which point the comparator 54 generates the oscillator control signal OSTOP at a logic low. As a result, the oscillator 60 deactivates or stops generation of the clock signals CLK and CLKN so as to stop ramp up of the charge pump 52.

This regulation scheme is effective, but must be set to account for the worst case operating scenario of the non-volatile memory so that the voltage of the output signal OUT does not result in breakdown of a parasitic diode, such as one within the triple well of a non-volatile memory. This requirement to account for the worst case operating scenario arises out of the fact that the regulation scheme does not track variations over process or temperature.

Therefore, further development is needed.

SUMMARY

A circuit disclosed herein includes a negative charge pump circuit configured to generate a negative charge pump output signal at a first node, the negative charge pump circuit being enabled as a function of at least one charge pump control signal. A diode has an anode coupled to the first node and a cathode coupled to a second node. The diode draws a reverse current from the second node. A comparator has a non-inverting input coupled to the second node and an inverting input coupled to a third node. The comparator is configured to cause generation of the at least one charge pump control signal. A current mirror arrangement is configured to source a first current to the second node and sink a second current from the third node.

The first current and second current may have a same magnitude.

The comparator may generate an oscillator control signal, and an oscillator may be configured to generate the at least one charge pump control signal as a function of the oscillator control signal generated by the comparator.

The circuit further may include a non-volatile memory, and the diode may be formed using a same process as transistors of the non-volatile memory.

The circuit is formed on a semiconductor substrate. A non-volatile memory is also formed in the semiconductor substrate. The diode is also formed in the semiconductor substrate.

The current mirror arrangement may include a current sink configured to sink the second current from the third node, and a current source configured to source the first current to the second node.

The current sink may include: a first NMOS transistor having a drain coupled to the third node, a source coupled to ground, and a gate; and a second NMOS transistor having a drain coupled to receive a first current, a source coupled to ground, and a gate coupled to the drain of the second NMOS transistor and to the gate of the first NMOS transistor.

The current source may include: a first PMOS transistor having a source coupled to a supply node, a drain coupled to the second node, and a gate; and a second PMOS transistor having a source coupled to the supply node, a drain coupled to the third node, and a gate coupled to the third node and to the gate of the first PMOS transistor.

The supply node may be coupled to a DC voltage supply or to a voltage regulator to receive a regulated voltage therefrom.

The voltage regulator may include: an operational amplifier having a non-inverting terminal coupled to receive a reference voltage, an inverting terminal, and an output providing the regulated voltage; and a voltage divider coupled between the output of the operational amplifier and ground, wherein a tap of the voltage divider is coupled to the inverting terminal of the operational amplifier.

The at least one charge pump control signal may include a clock signal and a complement of the clock signal.

Also disclosed herein is a circuit including a charge pump circuit generating a negative charge pump output signal at a first node as a function of at least one charge pump control signal, a diode having an anode coupled to the first node and a cathode coupled to a second node, a first current mirror configured to source a first current to the second node, and a second current mirror configured to sink a second current from a third node (the second current having a same magnitude as the first current). A comparison circuit compares a voltage at the second node to a voltage at the third node, causes the at least one charge pump control signal to direct the charge pump circuit to generate the negative charge pump output signal having a voltage that ramps downward, based upon the voltage at the second node being greater than the voltage at the third node; causes the at least one charge pump control signal to direct the charge pump circuit to cease the downward ramping of the voltage of the charge pump output signal, if the voltage at the second node is at least equal to the voltage at the third node.

A method disclosed herein includes generating a charge pump output signal that has a voltage that ramps downwardly, receiving the charge pump output signal at an anode of a diode, sourcing a second current to a cathode of the diode, sinking a first current from a node, the first current being equal in magnitude to the second current, comparing a voltage at the node to a voltage at the cathode of the diode. When voltage at the cathode of the diode becomes equal to the voltage at the node, the method continued with causing the charge pump output signal to cease ramping.

The charge pump output signal has a voltage that ramps upward in magnitude but is negative in sign. When the voltage at the cathode of the diode becomes equal to the voltage at the node, the charge pump output signal is caused to cease ramping upward in magnitude.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of a prior art charge pump and regulation circuit for the charge pump.

FIG. 2 is a detailed schematic diagram of a charge pump and regulation circuit for the charge pump, in accordance with this disclosure.

FIG. 3A is a block diagram of an integrated circuit including a non-volatile memory as well as a charge pump and regulation circuit such as that of FIG. 2 or FIG. 4.

FIG. 3B is a schematic block diagram of a system comprised of a non-volatile memory and a charge pump and regulation circuit such as that of FIG. 2 or FIG. 4.

FIG. 3C is a cross sectional view of transistors of the non-volatile memory of FIG. 3B.

FIG. 3D is a cross sectional view of a triple well CMOS pair.

FIG. 4 is a schematic diagram of a charge pump and regulation circuit for the charge pump, in accordance with this disclosure.

FIG. 5 is a schematic diagram of an alternative configuration of a charge pump and regulation circuit for the charge pump, in accordance with this disclosure.

FIG. 6 is a detailed schematic diagram of a charge pump and regulation circuit for a negative charge pump in accordance with this disclosure.

FIG. 7 is a detailed schematic diagram of another charge pump and regulation circuit for a negative charge pump in accordance with this disclosure.

DETAILED DESCRIPTION

The following disclosure enables a person skilled in the art to make and use the subject matter disclosed herein. The general principles described herein may be applied to embodiments and applications other than those detailed above without departing from the spirit and scope of this disclosure. This disclosure is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed or suggested herein. Throughout this disclosure, the term “coupled” is used, and means either a direct unbroken electrical connection between two components described as being coupled, or that is an indirect electrical connection between two components described as being coupled, such as though an intervening component.

Now described with reference to FIG. 2 is a charge pump and regulation circuit 100. The charge pump and regulation circuit 100 includes a charge pump 102 that generates a charge pump output signal 103 at node N1. The charge pump 102 is controlled by clock signals CLK and CLKN which serve as control signals for the charge pump 102 and are logical inverts of one another. The charge pump output signal 103 has a voltage of Vpump. A diode D1 has its cathode coupled to node N1 and its anode coupled to node N2.

A comparator 104 has its inverting input terminal coupled to node N2 and has its non-inverting input terminal coupled to the output of a current sourcing circuit 106 to receive a current I2. The comparator 104 generates an oscillator control signal STOP at its output. An oscillator 110 receives the oscillator control signal STOP and generates or outputs the clock signals CLK and CLKN based upon the oscillator control signal STOP, which act as an on/off enable signal for the oscillator 110. The oscillator control signal STOP is an on/off enable signal. When STOP is at a logic high, the oscillator 110 is enabled. When STOP is at a logic low, the oscillator 110 is disabled. A current sinking circuit 108 draws a current I3 from node N2. The currents I2 and I3 are equal in magnitude, and may each have a value of 1 μA, or different value depending on the design specifications.

A diode coupled NMOS transistor MN3 has its drain coupled to node N1, its source coupled to node Vout to produce output signal 105, and its gate coupled to its drain. The output signal 105 is provided to a non-volatile memory for use during program and erase options. A capacitor Cload represents capacitive load at node Vout. In some configurations, such as shown in FIG. 5, instead of a diode coupled NMOS transistor MN3, there is a diode coupled PMOS transistor MP3 that has its source coupled to node N1, its drain coupled to node Vout to produce output signal 105, and its gate coupled to its drain.

The current sourcing circuit 106 is comprised of a PMOS current mirror formed from PMOS transistors MP1 and MP2. PMOS transistor MP2 forms the input of the PMOS current mirror and has its source coupled to a supply node VDD, its drain coupled to a current source 107 that draws a current I1, and its gate coupled to its drain as well as to the gate of the PMOS transistor MP1. The PMOS transistor MP1 forms the output of the PMOS current mirror and has its source coupled to the supply node VDD and its drain coupled to node N3. The drain of the PMOS transistor MP1 sources current I2 to node N3, and the current I2 is equal to a multiple of the current I1, such as a multiple of 1 (meaning that current I2 may be equal to current I1). Therefore, PMOS transistor MP1 and PMOS transistor MP2 may be equal in size, or may be unequal in size, depending on the mirroring ratio of the current.

The current sinking circuit 108 is comprised of an NMOS current mirror formed from NMOS transistors MN1 and MN2. NMOS transistor MN2 forms the input of NMOS current mirror and has its drain coupled to node N3 to receive current I2, its source coupled to ground, and its gate coupled to its drain as well as to the gate of the NMOS transistor MN1 to make it a diode connected configuration. The NMOS transistor MN1 forms the output of the NMOS current mirror and has its drain coupled to the node N2 and its source coupled to ground. The drain of the NMOS transistor MN1 sinks current I3 from node N2, and current I3 is equal in magnitude to current I2. Current I3 is equal in magnitude to current I2 because the NMOS transistor MN1 is equal in size to the NMOS transistor MN2.

In operation, at startup, the voltage Vpart at node N2 is approximately at ground because little to no reverse current is flowing through the diode D1, while the current sink 108 is sinking current I3 from node N2. Since the current I2 is sourced by the current source 106 to the node N3, the voltage at the non-inverting terminal of the comparator 104 (node N3) will be greater than the voltage Vpart at node N2, which as explained, will be approximately at ground. Therefore, the oscillator control signal STOP from the comparator 104 remains at a logic high, enabling the oscillator 110, and the oscillator 110 will generate the clock signals CLK and CLKN to cause the charge pump 102 to ramp up the voltage Vpump of the charge pump output signal 103.

As Vpump begins to ramp up, the reverse current through the diode D1 begins to increase, but remains low. Until Vpump ramps up sufficiently for the reverse current through the diode D1 to be equal to the current I3 (which is known to occur for a given diode size at a voltage close to its breakdown voltage), the voltage Vpart at node N2 remains approximately at ground, maintaining the oscillator control signal STOP at a logic high, resulting ultimately in Vpump continuing to ramp up. Once Vpump ramps sufficiently for the reverse current through the diode D1 to be equal to the current I3, the voltage Vpart at the inverting terminal of the comparator 104 (node N2) will be equal to the voltage at the inverting terminal of the comparator 104 (node N3). At this point, the oscillator control signal STOP will be output by the comparator 104 at a logic low. As a result, the oscillator 110 will alter or stop the generation of the clock signals CLK and CLKN to cease ramp up of the voltage Vpump of the charge pump output signal 103 by the charge pump 104.

Since the oscillator control signal STOP will transition when the reverse current through the diode D1 is equal to the current I3, the maximum reverse current through the diode D1 is limited to the current I3 (which has the same magnitude as the current I2). Note that the reverse current through the diode D1 is a function of the actual reverse bias across the diode D1 in comparison to the actual breakdown voltage of the diode D1, meaning that the reverse current through the diode D1 increases as the current reverse bias of the diode D1 approaches the actual breakdown voltage of the diode Dl. This means that by monitoring the reverse current through the diode D1, it is known whether the actual reverse bias across the diode D1 is close to the actual breakdown voltage of the diode D1.

Therefore, by monitoring the reverse current and limiting it to a level known to be safe (below the breakdown voltage of the diode D1), the reverse bias across the diode D1 is inherently limited to a safe reverse bias. This is particularly useful because, provided that the charge pump and regulation circuit 100 is formed in the same integrated circuit substrate as a non-volatile memory to which the output signal 105 is provided, and provided diode D1 is formed to match the parasitic diodes within the triple well of floating gate MOS transistors within the non-volatile memory, by maintaining the diode D1 at a reverse bias that is known to be safe, those parasitic diodes within triple well of the floating gate MOS transistors of the non-volatile memory are maintained at a safe reverse bias, preventing damage to their respective floating gate MOS transistors.

Consequently, by monitoring the reverse current through the diode D1 and setting the voltage Vpump to cease ramping when the reverse current reaches a desired maximum level known to be safe, the breakdown voltage of the diode D1 (and thus the breakdown voltage of the parasitic diodes within the triple well of the non-volatile memory) is inherently tracked as it varies due to operating conditions, since Vpump will vary as the breakdown voltage of the diode D1 varies, maintaining the reverse bias across the diode D1 (and thus the reverse bias of the parasitic diodes within the floating gate MOS transistors of the non-volatile memory) at a safe level. In terms of voltage, the reverse bias across the diode D1 will be:
Vrb=Vpump−Vpart
where Vrb is the reverse bias across the diode D1 and Vpart is the voltage at node N2, which will be equal to the drain to source voltage of NMOS transistor MN1.

The voltage Vout of the output signal 105 produced at the source of the NMOS transistor MN3 will be:
Vout=Vpump−Vth
where Vth is the threshold voltage of NMOS transistor MN3.

As explained, provided that the charge pump and regulation circuit 100 is formed in the same integrated circuit substrate as a non-volatile memory to which the output signal 105 is provided, and provided diode D1 is formed to match the parasitic diodes within the floating gate MOS transistors within the non-volatile memory, by maintaining the diode D1 at a reverse bias that is known to be safe, those parasitic diodes within the floating gate MOS transistors of the non-volatile memory are maintained at a safe reverse bias, preventing damage to their respective floating gate MOS transistors. This arrangement is shown in FIG. 3A, where a memory chip 150 includes an integrated circuit substrate 152 in which the charge pump and regulation circuit 100 and a non-volatile memory 154 are formed. Note that therefore, the diode of the regulation circuit 100 and the parasitic diodes of the floating gate transistors of the non-volatile memory 154 are formed using a same process.

Shown in FIG. 3B is a schematic block diagram of the arrangement of the charge pump and regulation circuit 100 and non-volatile memory 154. Here, it can be seen that the charge pump and regulation circuit 100 generates the output signal 105 to a word line driver 201, n-well driver 202, and p-well driver 203. The word line driver 201 produces a control gate signal CG, the n-well driver 202 produces an n-well biasing signal NW, and the p-well driver 203 produces a p-well biasing signal PW. These signals 201, 202, and 203 are fed to memory transistors, as will be explained.

Depicted are first, second, and third triple well NMOS transistors T1, T2, and T3. The first triple well NMOS transistor T1 has its source coupled to a first bitline BL1, its drain coupled to the source of the second triple well NMOS transistor T2, its gate coupled to a selection signal SL, and its body coupled to the body of the second triple well NMOS transistor T2. The second triple well NMOS transistor T2 has its source coupled to the drain of the first triple well NMOS transistor T1 as stated, its drain coupled to a second bitline BL2, and its body coupled to the body of the first triple well NMOS transistor T1 as stated. The gate of the second triple well NMOS transistor T2 floats and is coupled to the gate of the third triple well NMOS transistor T3, which also floats. The third triple well NMOS transistor T3 has its source coupled to a control gate line CG, its gate coupled to the gate of the second triple well NMOS transistor T2 as stated, and its drain coupled to its body and its source.

The specific structure of the first, second, and third triple well NMOS transistors T1, T2, and T3 are shown in FIG. 3C. The source and drain of the first triple well NMOS transistor T1 are respectively formed by N+ regions 211 and 212 formed in a p-well 210. The source and drain of the second triple well NMOS transistor T2 are respectively formed by N+ regions 212 and 213 formed in the p-well 210, therefore note that the N+ region 212 is a source/drain region as it is both drain of the first triple well NMOS transistor T1 and source of the second triple well NMOS transistor T2. The gate structure 223 of the first triple well NMOS transistor T1 is formed between the N+ regions 211 and 212, and the gate structure 224 of the second triple well NMOS transistor T2 is formed between the N+ regions 212 and 213. A P+ region 214 used to for bulk biasing is formed in the p-well 210. The p-well 210 is surrounded by an epitaxial layer 216, and a buried N layer 220 underlies the epitaxial layer 220. Bordering the epitaxial layer 216 on both sides are respectively n-well 218 and n-well 221. N-well 218 has an N+ region 217 used for biasing the buried N layer 220 formed therein, and an N type isolation layer 219 is stacked between the n-well 218 and the buried N layer 220. An N type isolation layer 222 is stacked between the n-well 221 and the buried N layer 220.

The source and drain of the third triple well NMOS transistor T3 are formed by N+ regions 231 and 232 in a p-well 230. The gate structure 236 of the third triple well NMOS transistor T3 is formed between the N+ regions 231 and 232. A P+ region 233 used for bulk biasing is formed in the p-well 230. The p-well 230 is surrounded by an epitaxial layer 234, and the buried N layer 220 underlies the epitaxial layer 234. Bordering the epitaxial layer 234 on the side opposite to the n-well 221 is an n-well 235.

Note that a substrate 152 underlies the buried N layer 220, and that a P+ layer 241 used for substrate biasing is formed in the substrate 152.

As briefly explained above, parasitic diodes are inherent to the structure of MOS devices, due to the fact that a PN junction forms a PN junction diode. Therefore, due to the triple well structure of the NMOS structures T1, T2, and T3, in addition to the typical parasitic “body diodes” that would be present between the sources and drains, there are additional parasitic diodes. Parasitic diode D2 is formed between the n-well 218 and the substrate 152, while parasitic diode D3 is formed between the p-well 210 and buried N layer 220, and parasitic diode D4 is formed between the p-well 230 and the buried N layer 220. During program and erase operations, different voltage values on the control gate CG line, n-well biasing line NW, and p-well biasing line PW alter the biasing of the regions they are connected to. By tracking these parasitic diodes D2, D3, and D4 through the use of the diode D1 (see FIGS. 2 and 4 of the charge pump and regulation circuit 100), the output 105 of the charge pump and regulation circuit 100 is regulated to avoid breakdown of the parasitic diodes D2, D3, and D4.

It should be appreciated that the charge pump and regulation circuit 100 may also be used to help protect parasitic diodes within triple well CMOS designs, such as that shown in FIG. 3D. Here, a triple well NMOS transistor includes N+ regions 301 and 302 formed within a p-well 300. The gate structure 318 extends between the N+ regions 301 and 302. A P+region 303 used for bulk biasing is formed within the p-well 300. A deep n-well 304 underlies the p-well 300, and the p-well is bordered on its sides by the n-wells 305 and 307. A N+ region 306 used for biasing the deep n-well 304 is formed within the n-well 305, and likewise a N+ region 308 also usable for biasing the deep n-well 304 is formed within the n-well 307.

A triple well CMOS transistor includes P+ regions 311 and 312 formed within an n-well 309. The gate structure 319 extends between the P+ regions 311 and 312. An N+ region 310 used for bulk biasing is formed within the n-well 309. The n-well 309 is bordered by n-wells 307 and 313. The n-well 313 has an N+ region 314 formed therein also usable for biasing the deep n-well 304. A p-type substrate 315 underlies the deep n-well 304 and surrounds the n-wells 305 and 313. P+ regions 316 and 317 are formed within the p-type substrate 315 and are usable for biasing the p-type substrate.

In addition to the typical parasitic body diodes between source and drain, a parasitic diode D2 is formed between n-well 305 and the substrate 315, while a parasitic diode D3 is formed between the p-well 300 and the deep n-well 304, and a parasitic diode D4 is formed between the n-well 313 and the substrate 315.

By tracking these parasitic diodes D2, D3, and D4 through the use of the diode D1 (see FIGS. 2 and 4 of the charge pump and regulation circuit 100), the output 105 of the charge pump and regulation circuit 100 is regulated to avoid breakdown of the parasitic diodes D2, D3, and D4 due to different biases that may be generated based upon the output 105 and used to bias the p-well 300, substrate 315, and n-wells 305 and 313.

Owing to this regulation scheme, the voltage Vpump of the charge pump output 103 can be maintained at a level where the diode D1 is close to, but not at, its breakdown voltage. This provides for high efficiency in program and erase operations of the non-volatile memory 154 without fear of breakdown of the parasitic diodes within the triple well used for the non-volatile memory 154, because Vpump and therefore Vout can be maintained at a maximum safe level as operating conditions vary.

While this design eliminates the resistive divider of prior art designs, it adds components not present in those designs. There is therefore a net area gain of about 15%, however in addition there is a reduced cost of forming these designs, as high voltage rails are not necessary. Note that the area gain can vary with specific designs, applications, and specifications. In addition, this design permits any density of memory designs without impacting the efficiency of program and erase operations. Also, due to the lack of a resistive divider, the current load on the charge pump 102 is decreased, reducing power consumption by about 20%. This percentage may vary with specific designs, applications, and specifications, so power consumption reduction may be different in some instances.

It should be appreciated that the current sinking circuit 108 and current sourcing circuit 106 need not be comprised of current mirrors as shown in FIG. 2. Instead, as shown in the charge pump and regulation circuit 100 of FIG. 4, the current sinking circuit 108 may utilize any current source 109 that draws the current I3, and the current sourcing circuit 106 may utilize any current source 111 that provides the current I2, provided that the currents I3 and I2 are equal in magnitude.

Now described with reference to FIG. 6 is a charge pump and regulation circuit 200. The charge pump and regulation circuit 200 includes a negative charge pump 202 that generates a charge pump output signal 203 at node N1. The negative charge pump 202 is controlled by clock signals CLK and CLKN which serve as control signals for the negative charge pump 202 and are logical inverts of one another. The charge pump output signal 203 has a voltage of Vnegpump. A diode D1 has its anode coupled to node N1 and its cathode coupled to node N2.

A comparator 204 has its non-inverting input terminal coupled to node N2 and has its inverting input terminal coupled to the output of a current sinking circuit 206 to receive a current I2. The comparator 204 generates an oscillator control signal STOP at its output. An oscillator 210 receives the oscillator control signal STOP and generates or outputs the clock signals CLK and CLKN based upon the oscillator control signal STOP, which act as an on/off enable signal for the oscillator 210. The oscillator control signal STOP is an on/off enable signal. When STOP is at a logic high, the oscillator 210 is enabled. When STOP is at a logic low, the oscillator 210 is disabled. A current source circuit 208 sources a current I3 to node N2. The currents I2 and I3 are equal in magnitude, and may each have a value of 1 μA, or different value depending on the design specifications.

The output signal 205 (shown as Vout, formed across the load capacitance Cload) is provided to a non-volatile memory for use during program and erase options. Note that a current source 209 represents load current Iload sourced to current N1 as a result of the negative voltage Vnegpump (described in detail below) output by the negative charge pump 202.

The current sinking circuit 206 is comprised of an NMOS current mirror formed from NMOS transistors M3 and M4. NMOS transistor M4 forms the input of the NMOS current mirror and has its source coupled to a ground, its drain coupled to a current source 207 that draws a current I1 from VDD, and its gate coupled to its drain as well as to the gate of the NMOS transistor M3. The NMOS transistor M3 forms the output of the NMOS current mirror and has its source coupled to ground and its drain coupled to node N3. The drain of the NMOS transistor M3 sinks current I2 from node N3, and the current I2 is equal to a multiple of the current I1, such as a multiple of 1 (meaning that current I2 may be equal to current I1). Therefore, NMOS transistor M3 and NMOS transistor M4 may be equal in size, or may be unequal in size, depending on the mirroring ratio of the current.

The current sourcing circuit 208 is comprised of a PMOS current mirror formed from PMOS transistors M1 and M2. PMOS transistor M2 forms the input of PMOS current mirror and has its drain coupled to node N3 to receive current I2, its source coupled to VDD, and its gate coupled to its drain as well as to the gate of the PMOS transistor M1 to make it a diode connected configuration. The PMOS transistor M1 forms the output of the PMOS current mirror and has its drain coupled to the node N2 and its source coupled to VDD. The drain of the PMOS transistor M1 sources current I3 to node N2, and current I3 is equal in magnitude to current I2. Current I3 is equal in magnitude to current I2 because the PMOS transistor M1 is equal in size to the PMOS transistor M2.

In operation, at startup, the voltage Vpart at node N2 rises because little to no reverse current is flowing through the diode D1, while the current source 208 is sourcing current I3 to node N2. Since the current I2 is sunk by the current sink 206 from the node N3, the voltage at the non-inverting terminal of the comparator 204 (node N3) will be less than the voltage Vpart at node N2. Therefore, the oscillator control signal STOP from the comparator 204 remains at a logic high, enabling the oscillator 210, and the oscillator 210 will generate the clock signals CLK and CLKN to cause the negative charge pump 202 to ramp negatively downward the voltage Vnegpump of the charge pump output signal 203.

As Vnegpump begins to ramp down negatively (upwardly in magnitude, but negative in sign), the reverse current through the diode D1 begins to increase, but remains low. Until Vnegpump ramps down sufficiently for the reverse current through the diode D1 to be equal to the current I3 (which is known to occur for a given diode size at a voltage close to its breakdown voltage), the voltage Vpart at node N2 remains greater than the voltage at N3, maintaining the oscillator control signal STOP at a logic high, resulting ultimately in Vnegpump continuing to ramp negatively downward. Once Vnegpump ramps downward sufficiently for the reverse current through the diode D1 to be equal to the current I3, the voltage Vpart at the non-inverting terminal of the comparator 204 (node N2) will be equal to the voltage at the inverting terminal of the comparator 204 (node N3). At this point, the oscillator control signal STOP will be output by the comparator 204 at a logic low. As a result, the oscillator 210 will alter or stop the generation of the clock signals CLK and CLKN to cease the downward negative ramp of the voltage Vnegpump of the charge pump output signal 203 by the charge pump 204.

Since the oscillator control signal STOP will transition when the reverse current through the diode D1 is equal to the current I3, the maximum reverse current through the diode D1 is limited to the current I3 (which has the same magnitude as the current I2). Note that the reverse current through the diode D1 is a function of the actual reverse bias across the diode D1 in comparison to the actual breakdown voltage of the diode D1, meaning that the reverse current through the diode D1 increases as the current reverse bias of the diode D1 approaches the actual breakdown voltage of the diode D1. This means that by monitoring the reverse current through the diode D1, it is known whether the actual reverse bias across the diode D1 is close to the actual breakdown voltage of the diode D1.

Therefore, by monitoring the reverse current and limiting it to a level known to be safe (below the breakdown voltage of the diode D1), the reverse bias across the diode D1 is inherently limited to a safe reverse bias. This is particularly useful because, provided that the charge pump and regulation circuit 200 is formed in the same integrated circuit substrate as a non-volatile memory to which the output signal 205 is provided, and provided diode D1 is formed to match the parasitic diodes within the triple well of floating gate MOS transistors within the non-volatile memory, by maintaining the diode D1 at a reverse bias that is known to be safe, those parasitic diodes within triple well of the floating gate MOS transistors of the non-volatile memory are maintained at a safe reverse bias, preventing damage to their respective floating gate MOS transistors.

Consequently, by monitoring the reverse current through the diode D1 and setting the voltage Vnegpump to cease ramping negatively downward when the reverse current reaches a desired maximum level known to be safe, the breakdown voltage of the diode D1 (and thus the breakdown voltage of the parasitic diodes within the triple well of the non-volatile memory) is inherently tracked as it varies due to operating conditions, since Vnegpump will vary as the breakdown voltage of the diode D1 varies, maintaining the reverse bias across the diode D1 (and thus the reverse bias of the parasitic diodes within the floating gate MOS transistors of the non-volatile memory) at a safe level. In terms of voltage, the reverse bias across the diode D1 will be:
Vrb=Vpart−Vnegpump
where Vrb is the reverse bias across the diode D1 and Vpart is the voltage at node N2, which will be equal to the source to drain voltage of PMOS transistor M1.

In the embodiment of FIG. 6, the PMOS transistors M1 and M2 have their sources coupled to VDD, and current source 207 draws current I1 from VDD. However, in some cases, rather than VDD, the PMOS transistors M1 and M2 may draw a regulated voltage Vreg, and the current source 207 may draw current I1 from the regulated voltage Vreg. Such an embodiment of shown in FIG. 7, where the charge pump and regulation circuit 200′ includes a voltage regulator 502 that outputs the regulated voltage Vreg.

The voltage regulator 502 includes an operational amplifier 500 having its non-inverting terminal coupled to a reference voltage, and having its inverting terminal coupled to the center tap of a resistive divider formed by resistors R1 and R2 connected in series between the output of the operational amplifier 500 and ground. By using the voltage at this center tap as feedback, the operational amplifier 500 maintains the regulated voltage Vreg at a set steady level.

While the disclosure has been described with respect to a limited number of embodiments, those skilled in the art, having benefit of this disclosure, will appreciate that other embodiments can be envisioned that do not depart from the scope of the disclosure as disclosed herein. Accordingly, the scope of the disclosure shall be limited only by the attached claims.

Claims

1. A circuit, comprising:

a negative charge pump circuit configured to generate a negative charge pump output signal at a first node, the negative charge pump circuit being enabled as a function of at least one charge pump control signal;
a diode having an anode coupled to the first node and a cathode coupled to a second node, said diode drawing a reverse current from the second node, the reverse current drawn from the second node being a function of a reverse bias voltage across the diode;
a comparator having a non-inverting input coupled to the second node and an inverting input coupled to a third node, the comparator configured to cause generation of the at least one charge pump control signal; and
a circuit arrangement configured to source a first current to the second node and sink a second current from the third node.

2. The circuit of claim 1, wherein the first current and second current have a same magnitude.

3. The circuit of claim 1, wherein the comparator generates an oscillator control signal; and further comprising an oscillator configured to generate the at least one charge pump control signal as a function of the oscillator control signal generated by the comparator.

4. The circuit of claim 1, wherein the circuit further comprises a non-volatile memory; and wherein the diode is formed using a same process as transistors of the non-volatile memory.

5. The circuit of claim 1, wherein the circuit comprises a semiconductor substrate; further comprising a non-volatile memory formed in the semiconductor substrate; and wherein the diode is also formed in the semiconductor substrate.

6. The circuit of claim 1, wherein the circuit arrangement comprises:

a current sink configured to sink the second current from the third node; and
a current source configured to source the first current to the second node.

7. The circuit of claim 6, wherein the current sink comprises:

a first NMOS transistor having a drain coupled to the third node, a source coupled to ground, and a gate; and
a second NMOS transistor having a drain coupled to receive a first current, a source coupled to ground, and a gate coupled to the drain of the second NMOS transistor and to the gate of the first NMOS transistor.

8. The circuit of claim 7, further comprising a current source configured to draw the first current from a supply node and supply the first current to the drain of the second NMOS transistor.

9. The circuit of claim 6, wherein the current source comprises:

a first PMOS transistor having a source coupled to a supply node, a drain coupled to the second node, and a gate; and
a second PMOS transistor having a source coupled to the supply node, a drain coupled to the third node, and a gate coupled to the third node and to the gate of the first PMOS transistor.

10. The circuit of claim 9, wherein the supply node is coupled to a DC voltage supply.

11. The circuit of claim 9, wherein the supply node is coupled to a voltage regulator to receive a regulated voltage therefrom.

12. The circuit of claim 11, wherein the voltage regulator comprises:

an operational amplifier having a non-inverting terminal coupled to receive a reference voltage, an inverting terminal, and an output providing the regulated voltage;
a voltage divider coupled between the output of the operational amplifier and ground, wherein a tap of the voltage divider is coupled to the inverting terminal of the operational amplifier.

13. The circuit of claim 1, wherein the at least one charge pump control signal comprises a clock signal and a complement of the clock signal.

14. A circuit, comprising:

a charge pump circuit generating a negative charge pump output signal at a first node as a function of at least one charge pump control signal;
a diode having an anode coupled to the first node and a cathode coupled to a second node;
a first circuit configured to source a first current to the second node;
a second circuit configured to sink a second current from a third node, the second current having a same magnitude as the first current; and
a comparison circuit configured to: compare a voltage at the second node to a voltage at the third node; cause the at least one charge pump control signal to direct the charge pump circuit to generate the negative charge pump output signal having a voltage that ramps downward, based upon the voltage at the second node being greater than the voltage at the third node; and cause the at least one charge pump control signal to direct the charge pump circuit to cease the downward ramping of the voltage of the charge pump output signal, if the voltage at the second node is at least equal to the voltage at the third node.

15. The circuit of claim 14, wherein the comparison circuit is configured to:

cause the at least one charge pump control signal to direct the charge pump circuit to generate the charge pump output signal having a voltage that ramps downward, if the voltage at the second node is greater than the voltage at the third node; and
cause the at least one charge pump control signal to direct the charge pump circuit to cease downward ramp of the voltage of the charge pump output signal, if the voltage at the second node is at least equal to the voltage at the third node.

16. The circuit of claim 14, wherein the first circuit comprises a first current mirror having an input coupled to the third node and an output coupled to the second node.

17. The circuit of claim 16, wherein the second circuit comprises a second current mirror has an input coupled to a current source and an output coupled to the third node.

18. The circuit of claim 14, wherein the circuit further comprises a non-volatile memory; and wherein the diode is formed using a same process as transistors of the non-volatile memory.

19. The circuit of claim 14, wherein the circuit comprises a semiconductor substrate; further comprising a non-volatile memory formed in the semiconductor substrate; and wherein the diode is also formed in the semiconductor substrate.

20. A method, comprising:

generating a charge pump output signal that has a voltage that ramps downwardly;
receiving the charge pump output signal at an anode of a diode;
sourcing a second current to a cathode of the diode;
sinking a first current from a node, the first current being equal in magnitude to the second current;
comparing a voltage at the node to a voltage at the cathode of the diode; and
when the voltage at the cathode of the diode becomes equal to the voltage at the node, causing the charge pump output signal to cease ramping.

21. The method of claim 20, wherein the charge pump output signal has a voltage that ramps upward in magnitude but is negative in sign; wherein, when the voltage at the cathode of the diode becomes equal to the voltage at the node, the charge pump output signal is caused to cease ramping upward in magnitude.

22. The method of claim 20, wherein sourcing the second current to the cathode of the diode is performed by mirroring the first current to be drawn from node.

23. The method of claim 20, wherein sinking the first current from the node is performed by mirroring a reference current to the node.

Referenced Cited
U.S. Patent Documents
5553295 September 3, 1996 Pantelakis
5692164 November 25, 1997 Pantelakis
5721704 February 24, 1998 Morton
5986947 November 16, 1999 Choi et al.
6418040 July 9, 2002 Meng
6445243 September 3, 2002 Myono
7145370 December 5, 2006 Bernard et al.
7342438 March 11, 2008 Muneer et al.
7427890 September 23, 2008 Chan
7741898 June 22, 2010 Hsu
8179734 May 15, 2012 Soma
8693256 April 8, 2014 Pasotti et al.
9159425 October 13, 2015 Rana et al.
9627441 April 18, 2017 Van Buskirk
20080157729 July 3, 2008 Chan
20110002164 January 6, 2011 Soma
20130010536 January 10, 2013 Wang et al.
20170063369 March 2, 2017 Boakye et al.
20190213234 July 11, 2019 Bayat et al.
Foreign Patent Documents
1627448 June 2005 CN
101894586 November 2010 CN
102354521 February 2012 CN
107924205 April 2018 CN
Other references
  • Soldera, J., et al: “A Low Ripple Fully Integrated Charge Pump Regulator,” Proceedings of the 16th Symposium on Integrated Circuits and Systems Design, (SBCCI'03), 2003 IEEE (4 pages).
  • Calligaro, Cristiano, et al: “Voltage Multiplier With Output Stabilization,” 1996 IEEE (pp. 905-908).
  • Kim, Sung-Eun, et al: “A Small Ripple Regulated Charge Pump With Automatic Pumping Control Schemes,” 2004 EEE (pp. 383-386).
  • Shor, Joseph S., et al: “Self-Regulated Four-Phased Charge Pump with Boosted Wells,” 2003 IEEE (pp. I-241 thru I-244).
  • First Office Action and Search Report for co-pending CN Appl. No. 202010066352.2 dated Jun. 4, 2021 (6 pages).
Patent History
Patent number: 11258358
Type: Grant
Filed: Jan 14, 2020
Date of Patent: Feb 22, 2022
Patent Publication Number: 20200235660
Assignee: STMicroelectronics International N.V. (Schiphol)
Inventors: Vikas Rana (Noida), Shivam Kalla (Bikaner)
Primary Examiner: Patrick C Chen
Application Number: 16/742,248
Classifications
Current U.S. Class: Including Semiconductor Means (363/60)
International Classification: H02M 3/07 (20060101);